Marielle's little place on the web

In role playing games, players are faced with the task of creating a role playing character. For good role playing, these characters need a detailed background story. They have a personality, their own wishes, their own quirks and their own appearance. Many tricks exist to get to exciting characters fast. Personas are very much like role playing characters and the same techniques can be used to create interesting personas. continue reading…

Applying results from studies to other contexts is not easy and should always be done with caution. I came across yet another example, this time someone telling that human behavior in lab settings can be very different from human behavior in a natural environment. If one wants to know more about natural behavior, one should study this behavior as it naturally occurs. With the current technological means, this can be very easy (video surveillance, gps tracking, logging internet activity). And yes, there is no doubt that the data acquired in this way describes the behavior one wanted to study. But why then test in a lab setting?

The main disadvantage of lab experiments is that they take place in an artificial environment, with non-natural tasks and non-natural conditions. In the same time, this is a big advantage. Lab experiments can be reproduced, every detail can be carefully controlled and the results are easier to interpret because there are fewer unknown factors. Although the results may not tell much about human behavior in natural settings, they do tell a lot about human behavior in lab settings, which valuable information in itself. Using this information to speculate or hypothesize about behavior in other settings is perfectly fine. Drawing conclusions is not.

GMI, a company in market research, offers a product for click testing, which according to their claims can be used to ‘[accurately measure] what consumers notice on the [web] page’. Test subjects are shown a couple of pages and are asked for each page to click on the 5 elements that first grab their attention. The software uses these data to generate a so-called ‘heat map’.

It’s nice software to play with, and the heat maps it generates are certainly appealing and interesting. A little critical note though. We know from cognitive psychology that there can be great differences between what users think and what they claim they think. This is one of the reasons why ‘thinking out loud’-tests have only limited value. GMI’s software does not measure which elements grab the user’s attention, but can provide valuable insights on what users think they focus on, in an easy manner and at relatively low cost.